VC Capital Shifts Toward Defense AI and Hardware Infrastructure

The striking part is not the headline number. It is where the money went: companies with fielded hardware, enterprise customers, or both—rather than another cohort of software demos hunting for a business model.

For LPs, this is a familiar trade: higher technical and regulatory friction in exchange for a closer line of sight to procurement, deployment and revenue. The burn rate did not disappear. It just acquired an export-control file.

Defense AI is becoming a capital concentration trade

The report puts disclosed defense-AI funding at $3 billion across two German companies tied to the NATO supply chain.

Helsing raised a $1.8 billion Series E at an $18 billion post-money valuation, according to the report. Quantum Systems raised $1.2 billion in a Series D at an approximately $8 billion valuation. The former develops AI software for autonomous weapons systems; the latter develops autonomous aerial vehicles and a multi-domain autonomy platform.

That is not broad-based venture diversification. It is concentration at scale.

• Two companies accounted for the full disclosed defense-AI total cited in the report.
• Both operate in a category where government procurement can create real commercial pathways—and where contract cycles, compliance and jurisdictional exposure can also rewrite an investment case.
• The report flags export-control exposure around AI-enabled autonomous weapons, including US-origin technology rules that may apply even when the funded company is European.

The market narrative says “European sovereignty.” The underwriting question is less poetic: who controls the customer relationship, the technology stack and the permissions to sell across borders?

Lakestar’s Resilience I close adds another data point. EU-Startups reports that the firm closed the €262.2 million fund, equivalent to $300 million, to invest in European dual-use and defense technology startups. The category has moved from awkward allocation conversation to dedicated institutional vehicle. That does not make it liquid.

Quantum just got a much larger price tag

London-based quantum hardware company Oratomic closed a $300 million Series A co-led by ARCH Venture Partners, Spark Capital and Khosla Ventures, according to The Innovation Attorney. The report describes it as a record quantum venture round by size.

A $300 million A round is a blunt signal: investors are funding the hardware buildout before the category has the luxury of cheap iteration. Quantum hardware needs capital for systems, talent and development—not merely cloud credits and a sharp pitch deck.

The upside case is obvious enough. The less glamorous version is that a large first institutional round can turn follow-on financing into a much tougher hurdle. Bigger cap tables and bigger operating requirements raise the bar for the next proof point. For investors entering later, price discipline matters more than thematic enthusiasm.

Physical deployment is the common denominator

The report says physical robotics drew five financing events totaling more than $350 million during the week, while agentic AI platforms for regulated financial, legal and healthcare workflows also drew five rounds totaling more than $350 million. Stablecoin infrastructure and AI-assisted drug discovery each produced a transaction above $100 million.

The common thread is deployment pressure. Robotics must work on construction and industrial sites. Regulated workflow software must survive the compliance environment it claims to streamline. Defense systems must clear procurement and export constraints.

This is not a return to venture normality. It is a rotation toward businesses where the demo is the easy part.

What this actually means for LPs: capital is flowing toward technically defensible, operationally heavy assets with longer underwriting chains. The diligence list should follow: customer concentration, procurement dependency, export restrictions, financing runway and follow-on reserves. “Mission critical” remains a compelling pitch. It is not, by itself, a liquidity event.